INTEREST IN THE QUALITY of condoms continues to be high because of their role in helping to prevent the spread of acquired immunodeficiency syndrome and other sexually transmitted diseases.
Part of the concern about condom quality arises because the water leak test used to ensure quality of condoms cannot detect some small holes that might allow small amounts of virus penetration during use. It has been estimated that in ideal circumstances the water leak test can detect holes down to 3 to 4 μm in diameter.1,2 Sexually transmitted viruses, on the other hand, with diameters of 0.04 to 0.15 μm,3 are smaller than these holes by more than an order of magnitude. If relative size were the only important issue, one might deduce that because the quality-control test does not detect very small holes, some condoms may pass the water leak test and in theory still allow some amount of virus penetration in actual use. However, the apparent insensitivity of the quality assurance test does not mean that many, if any, condoms have holes that might allow virus passage. Furthermore, virus transmission through a small hole depends on several factors in addition to hole size, including challenge virus titer, transcondom pressure, time for passage, viscosity of the carrier fluid, and condom thickness.4 This leaves the question of just how good as virus barriers are condoms that are available to the consumer?
Some data are available on the quality of off-the-shelf latex condoms as virus barriers. One laboratory barrier test used some physiologic-based conditions (e.g., transcondom peak pressure [60 mm Hg], pH, salinity, temperature, challenge particle size) and some exaggerated conditions (e.g., lower viscosity, high average pressure, extended test duration [30 minutes])4 and found that a proportion (29 of 89 samples) of representative latex condoms allowed small amounts of penetration by human immunodeficiency virus (HIV)-size polystyrene spheres.5 These results have been quoted widely, often out of context, resulting in the effectiveness of condoms as virus barriers being called into question. At the same time, however, a similar laboratory test, based on the Retta method4 but using a small challenge virus, φX174 (27 nm diameter), instead of the polystyrene spheres, independently found that only a few percent of latex condoms (3 of 60 samples) allowed some penetration.6
The purpose of the current study was to use a sensitive, conservative test to evaluate a broader range of condom types and brands as barriers to the small virus challenge to determine the proportion of condoms on the market that might allow virus penetration and the amount of virus penetration each allowed.
Materials and Methods
The challenge virus used in this study was the bacteriophage φX174 (27-nm particle diameter; 32 nm, including bulky “spikes”).7 It was used as a surrogate3 for pathogenic human viruses primarily because it can be used with greater safety and at much less cost and because its small size allows a conservative test. In addition, it is the least adsorbing virus of all those tested to date,8–10 and thus least likely to lead to artifacts resulting from virus adsorption in buffer.10,11 The host bacterium for the bacteriophage and the methods of preparation and assay have been published.12 The challenge buffer, Dulbecco's phosphate-buffered saline (DPBS), included 0.1% Triton X-100 (Sigma Chemical Co., St. Louis, MO), a nonionic (polar) biologic detergent, to reduce the surface tension.6 The challenge buffer contained 1 to 3 × 108 pfu/ml of the bacteriophage.
Male latex (9 brands, 470 samples) and polyurethane (2 brands, 76 samples) condoms were obtained through retail distributors (drug stores and grocery stores). They were not pretested in our laboratory for conformance to quality assurance tests.
Some of the condoms were lubricated (with or without nonoxynol-9 spermicide). Before testing, most of the lubrication was gently removed by rinsing with DPBS and blotting with sterile paper towels. No other manipulation was done so that the test evaluated condom permeability immediately out of the package.
A new test apparatus was designed to improve on the simple, inexpensive apparatus originally developed in this laboratory.6 The new apparatus provided a more sensitive test and reduced the possibilities of condom damage or occasional airborne contamination. The details of the apparatus are described elsewhere.13,14
Condom Test Protocol
The test protocol was essentially the same as used previously,6 with minor modifications.14 Briefly, the condom was fitted onto the test apparatus and kept from expanding under pressure beyond anthropomorphic dimensions by an open-mesh restrainer (made of nylon organdy or polyester organza material), was filled with a challenge virus-containing buffer (DPBS, pH 7) plus surfactant (0.1% Triton X-100) under 60 mm Hg (32 inches of water) constant hydrostatic pressure, and was immersed in the buffer (DPBS, without Triton X-100) to collect any virus particles that penetrated the condom. The collection buffer was then assayed at 30 minutes to detect the viruses.
The amount of virus penetration was calculated as the amount of challenge virus suspension needed to account for the amount of virus found in the collection buffer (i.e., amount of virus in the collection buffer divided by the titer of the challenge, presented in milliliters). With a challenge titer of 108 pfu/ml, the test could detect virus penetration of 2 × 10−6 ml.
Experiments using latex condoms with laser-produced holes of known size14 indicated that this protocol can reliably detect φX174 penetration through holes of 1.5 to 2.5 μm in diameter (14 of 16 samples), with an apparent limit of 1 to 1.5 μm (beyond which no virus penetration would be detectable). This experimental limit in detectability occurs because of the severely limited flow of fluid through a hole so small. In addition, it was found that essentially the same results were obtained with restrainers made of either type of open-mesh material (nylon organdy or polyester organza).
As reported for previous tests of virus penetration of latex condoms,6,11,14 controls were used to ensure against false-negative or false-positive results. Preliminary investigation demonstrated that virus present in the challenge buffer (at high titer) or in the collection buffer (at low titer) did not decrease over the duration of the test. Visual detection of leaks in the test apparatus (e.g., at tubing or valve connections) and open agar plates for aerosol detection were used to ensure against falsepositive results.
The results from the tests of the latex condoms are presented in Table 1 and Figure 1. Twelve of the 470 condoms (2.6%) allowed some amount of virus penetration. Data are presented in the figure as the percentage of condoms that allowed penetration of the indicated volume (or more) of challenge virus suspension.
Results for lubricated condoms are shown in Table 2. The data indicate that the presence of lubrication did not have a significant effect on the percentage of condoms that allowed virus penetration. Thus, the lubrication and rinsing did not plug or close many, if any, existing holes, and the handling of the condoms during the process of removing the lubrication did not increase virus penetration, or the two processes balanced the effects of each other.
The two materials used to construct the restrainers yielded similar results: for organdy, 6 of 253 (2.4%) condoms allowed virus penetration; for organza, 6 of 217 (2.8%) allowed penetration.
Data on penetration of φX174 through condoms made of polyurethane are also shown in Table 1 and Figure 1. Four of the 76 condoms (5.3%) allowed virus penetration. These results are not statistically different from those for latex condoms by the two-sided Fisher's exact test (P = 0.3).
Condoms as Barriers to Viruses
These results demonstrated that few latex condoms allowed virus penetration in a sensitive test of barrier integrity with a small virus (i.e., few condoms have holes large enough or numerous enough to allow detectable virus penetration). This suggests that the basic latex membrane was normally impervious to virus passage and that holes occurred infrequently.
The “failure rate” in this laboratory test was 2.6% (12 of 470). Most (99.8%) of the virus penetration came from just two condoms, so care must be used in presenting the overall virus penetration. Among the “failures,” the median penetration value might be considered representative: 7 × 10−4 ml.
The proportion of latex condoms that allowed penetration of the small virus in this study (12 of 470) was consistent with that found earlier in this laboratory (3 failures among 60 condoms)6 and was lower than that reported previously by Carey et al5 for penetration of HIV-sized polystyrene spheres. However, in their case, although 29 of 89 condoms allowed particle penetration, the average amount was only 0.0016 ml/condom (Carey, personal communication). Thus, although superficially the results appear to disagree, both sets of data indicate a high degree of barrier effectiveness under the different test conditions.
The results with the polyurethane condoms were similar to those for the latex condoms. The median amount of virus penetration was about 4 × 10−3 ml. Again, of the few condoms that allowed virus penetration, one condom accounted for 98.6% of the total amount.
Extrapolation to Actual Use Predictions
The amounts of virus penetration determined by this experimental protocol represent an exaggeration of what to expect from these holes in actual use. Several of the parameters of the test used physiologic-based conditions, and some conditions were exaggerated to provide a conservative test of condom barrier effectiveness. To extrapolate from these data to expected actual use values, correction values suggested by Carey et al5 can be used for viscosity (0.07, for semen) and average pressure (0.2), giving a net correction value of 0.014. Applying this value to the median value of virus penetration for the latex condoms that allowed penetration yields a representative value for penetration of semen in actual use: 1 × 10−5 ml. The condom that allowed the greatest amount of virus penetration represented a worst case of about 0.05 ml (3.5 ml × 0.014).
Consideration of condom effectiveness against specific pathogenic viruses (e.g., HIV or hepatitis B virus) requires further extrapolation to include actual virus titers and infectivities and is not dealt with in this report. Further, this study has focused on the initial holes in condoms. An assessment of risk associated with condom use should also include changes in holes and condom breakage during use.
In summary, this evidence indicates that although latex and polyurethane condoms are not perfect, they are substantial barriers to virus transmission.
1. Herman BA, Carey RF, Rinaldi JE. Sensitivity of water leak tests for latex condoms. J Test Eval 1993; 21:124–128.
2. Schmukler R, Casamento J, Baier RE, Beard RB. Testing of the barrier function of condoms: An overview. Presented at the Annual Meeting of the American Society of Mechanical Engineers, San Francisco, California, December 10–15, 1989.
3. Lytle CD, Truscott W, Budacz AP, Venegas L, Routson LB, Cyr WH. Important factors for testing barrier materials with surrogate viruses. Appl Environ Microbiol 1991; 7:2549–2554.
4. Retta SM, Herman WA, Rinaldi JE, Carey RF, Herman BA, Athey TW. Test method for evaluating the permeability of intact prophylactics to viral-size microspheres under simulated physiologic conditions. Sex Transm Dis 1991; 18:111–118.
5. Carey RF, Herman WA, Retta RS, Rinaldi JE, Herman BA, Athey TW. Effectiveness of latex condoms as a barrier to human immunodeficiency virus-sized particles under conditions of simulated use. Sex Transm Dis 1992; 19:230–234.
6. Lytle CD, Routson LB, Cyr WH. A simple method to test condoms for penetration by viruses. Appl Environ Microbiol 1992; 58:3180–3182.
7. McKenna R, Xia D, Willingmann P, et al. Atomic structure of single-stranded DNA bacteriophage fX174 and its functional implications. Nature 1992; 355:137–143.
8. Shields PA. Factors influencing virus adsorption to solids. PhD Dissertation, University of Florida, Gainesville, Florida, 1986.
9. Shields PA, Farrah SR. Determination of the electrostatic and hydrophobic character of enteroviruses and bacteriophages (Abstract Q-82). In: Program of Abstracts for the Annual Meeting of the American Society for Microbiology. Washington, DC: American Society for Microbiology, 1987.
10. Lytle CD, Routson LB. Minimized virus binding for tests of barrier materials. Appl Environ Microbiol 1995: 61:643–649.
11. Lytle CD, Routson LB, Thomas DP, Regnault WF, Cyr WH. Two parameters limiting the sensitivity of laboratory tests of condoms as viral barriers. J Test Eval 1996: 24:279–286.
12. Lytle CD, Budacz AP, Keville E, Miller SA, Prodouz KN. Differential inactivation of surrogate viruses with merocyanine 540. Photochem Photobiol 1991: 54:489–493.
13. Lytle CD, Seaborn GB, Routson LB, Dixon LG, Thomas DP, Cyr WH. Limits on laboratory testing of condoms as viral barriers. Presented at the Association of Official and Analytical Chemists International Meeting, Nashville, Tennessee, September 17–21, 1995.
14. Lytle CD, Routson LB, Duff JE, Fleharty B, Cyr HW. A sensitive method to evaluate condoms as virus barriers. J AOAC 1997 (in press).